Workshop

Download course materials from here.

bit.ly/———–

Basic ggplot2

# https://bookdown.org/ndphillips/YaRrr/pirateplot.html
library(tidyverse)
install.packages("ggbeeswarm")
library(ggbeeswarm)

Exploration

Plot boxplots and violin plots for the ratings. Subset by participant.

data %>%
  filter(region == 1) %>%
  ggplot(aes(x=gram, y=rating, fill=freq)) +
  geom_boxplot(position = position_dodge(.9)) +
  #geom_violin(alpha=.5) +
  facet_wrap(~subj)

Group by region, word, frequency, and grammaticality. Summarise mean and standard error.

data %>%
  group_by(region, word, freq, gram) %>%
  summarise(meanRT = mean(rt),
            seRT = sd(rt)/sqrt(n())) %>%
  ggplot(aes(x=region,y=meanRT,colour=gram)) +
    geom_point() +
    geom_path(aes(lty=freq)) +
    geom_errorbar(aes(ymin=meanRT-seRT, ymax=meanRT+seRT),width=.1) +
  scale_x_continuous(breaks=c(1:5),labels = c("the","old","VERB","the","boat")) +
  ylab("reading time (miliseconds)")

Why?

Well, check out the Datasaurus Dozen, which all have the same x/y means and standard deviations:

Try it out

Dataset iris:

The function ggplot layers different geometries and aesthetics to build up a plot:

iris %>%
  mutate(Sepal.Width = Sepal.Width+rnorm(length(Sepal.Width),0,.1))%>%
  ggplot(aes(x=Species,y=Sepal.Width,fill=Species)) +
  geom_violin(lty=0,alpha=.5)+
  geom_boxplot(alpha=0.5,lwd=.5) +
  geom_quasirandom(dodge.width=1, alpha=.5)

In what ways might we change this plot?

iris %>%
  mutate(
    Sepal.Width = Sepal.Width + 
      rnorm(
        length(Sepal.Width) , 0 , .1
        )
    ) %>%
  ggplot(aes(x=Species,y=Sepal.Width)) +
  geom_boxplot() +
  geom_quasirandom(aes(pch=Species), alpha=.5)
  #ggplot(aes(x=Species,y=Sepal.Width,fill=Species)) +
  #geom_violin(lty=0,alpha=.5)+
  #geom_boxplot(alpha=0.5,lwd=.5) +
  #geom_quasirandom(dodge.width=1, alpha=.5)

Going back to quakes, let’s pipe our table into a ggplot (and fill in missing values):

quakes %>%
  group_by(mag) %>%
  summarise(n=n(),
            sta=mean(stations),
            staSD=sd(stations),
            staSE=staSD/sqrt(n),
            dep=mean(depth),
            depSD=sd(depth),
            depSE=depSD/sqrt(n)) %>%
  ggplot(aes(x=mag)) +
  geom_point(aes(y=sta),colour="red")+
  geom_path(aes(y=sta),colour="red")+
  geom_ribbon(aes(ymin=sta-staSD,ymax=sta+staSD),fill="red",alpha=.2) +
  geom_ribbon(aes(ymin=sta-staSE,ymax=sta+staSE),fill="red",alpha=.5) +
  geom_point(aes(y=dep),colour="blue")+
  geom_path(aes(y=dep),colour="blue")+
  geom_ribbon(aes(ymin=dep-depSD,ymax=dep+depSD),fill="blue",alpha=.2) +
  geom_ribbon(aes(ymin=dep-depSE,ymax=dep+depSE),fill="blue",alpha=.5) +
  theme_bw() + ylab("depth OR number of stations reporting")

Inheritance:

ggplot(data=quakes, aes(x=long)) +
  geom_point(aes(y=lat,colour=stations)) #+
  #geom_path()